Text
                    AMC PAMPHLET	АМСР 706-248
(THIS IS A REPRINT OF ORDP 20-248, WHICH WAS REDESIGNATED AMCP 706-248.)
ENGINEERING DESIGN
HANDBOOK
AMMUNITION SERIES
SECTION 5
INSPECTION ASPECTS OF ARTILLERY
AMMUNITION DESIGN
STATEMENT -2 UNCLASSIFIED
Tills document is subject to special export
c on t rtds -ind each t г i ns m it t a i to foreign
gOVl i iiin nt -
nnlv u- It ’ pt
(' ammarr: .	' t
or foreign nationals may be made
Lor approval of: Army Materiel
tn AMCRD-TV, Washington, D.C.
20 t! 5
шлицилп I L?.J,
U.0. nnmi IHNliniLL uu MMAND
MARCH 1966

BLANK PAGE
HEADQUARTERS UNITED STATES ARMY MATERIEL COMMAND WASHINGTON, D. C. 20315 10 March 1966 AMCP 706-248, Section Inspection Aspects of Artillery Ammu- nition Design, forming part of the Ammunition Series of the Army Materiel Command Engineering Design Handbook Series, is published for the information and guidance of all concerned. (AMCRD) FOR THE COMMANDER: г;' < J; ' i OFFICIAL: rSAWI CKI i Qi . ( У 9* Colonel, GS 1 //1 -I 9 Chief, Administrative bffic^ | DISTRIBUTION: Special 4 SELWYN D. SMITH, JR. * Major General, USA Chief of Staff
PREFACE The Engineering Design Handbook Series of the Army Materiel Command is a coordinated series of handbooks containing basic in- formation and fundamental data useful in the design and develop- ment of Army materiel and systems. The handbooks are authorita- tive reference books of practical information and quantitative facts helpful in the design and development of Army materiel so that it will meet the tactical and the technical needs of the Armed Forces. This handbook 1s the fifth of six handbooks on artillery ammunition and forms a part of the Engineering Design Handbook Series of the Army Materiel Command. Information concerning the other handbooks on artillery ammunition, together with the Table of Contents, Glossary and Index, will be found in AMCP 706-244, Section lt Artillery Ammunition—General. The material for this series was prepared by the Technical Writing Service of the McGraw-HillBook Co., based on technical information and data furnished principally by Picatinny Arsenal. Final preparation for publication was accomplished by the Engi- neering Handbook Office of Duke University, Prime Contractor to the Army Research Office-Durham for the Engineering Design Hand- book Series. Elements of the U. S. Army Materiel Command having need for handbooks may submit requisitions or official requests directly to Publications and Reproduction Agency, Letterkenny Army Depot, Chambersburg, Pennsylvania 17201. Contractors should submit such requisitions or requests to their contracting officers. Comments and suggestions on this handbook are welcome and should be addressed to Army Research Office-Durham, Box CM, Duke Station, Durham, North Carolina 27706. 1
TABLE OF CONTENTS lection I - Incpoctlon Aspect* off Artillery Ammunition Do*i«n Page Paragraph Quality Auurance Aspects of Ammunition Design 5-1 Quality Assurance 5-1 5-1 Amount of Inspection 5-1 5-2 Definition of Lot* 5-1 5-3 Sampling Risks 5-4 Operating Characteristic Curve i 5-5 The Average Outgoing Quality (AOQ) 5-3 5-6 Establishing the Acceptable Quality Level (AQL) 5-4 5-7 Classification of Defect* 5-5 5-8 Inspection by Attributes 5-5 5-9 Single-Sampling 5-5 5-10 Double-Sampling 5-5 5-11 Multiple-Sampling 5-6 5-12 ОС Curve* for Comparable Single-, Double-, and Multiple-Sampling Plans 5-6 5-13 Relationship of Sample Site to Lot Sire 5-6 5-14 Acceptable Quality Level as Basis for Inspection 5-8 5-15 Resubmission and Retest 5-10 5-16 Continuous-Sampling Plans 5-10 5-17 Statistical Aspect of Parallel Design 5-11 5-18 Relationship Between Sampling Plan, Tolerance Limits, and Safety Factor 5-11 5-19 Sampling Plans Based on Variables 5-12 5-20 Variables Inspection Compared with Attribute Inspection 5-12 5-21 References and Bibliography 5-12 Effect of Dimensioning and Tolerancing on Inspection 5-13 Introduction 5-13 5-22 Locational Tolerance Symbols 5-13 5-23 Independent Locational Tolerance 5-13 5-24 Dependent Locational Tolerance 5-15 5-25 Centrality of Holev 5-22 5-26 Basic Angle Dimensioning 5-23 5-27 Effect of Gage Tolerance on Component Tolerance 5-24 5-28
SECTION К INSPECTION d ASPECTS OF ARTILLERY AMMUNITION DESIGN QUALITY ASSURANCE ASPECTS OF AMMUNITION DESIGN 5-1. Quality Assurance It is necessary not only to state the dimensions to which an item must be produced, and the nature and propertigs of the materials of which the item must be made, but also to state methods for determining whether these requirements have been met to an extent which will be satisfactory to the Gov- ernment. The term "quality assurance" embraces the techniques used in the determination of the ac- ceptability of products. Th-'se techniques in- clude*. 1.* Establishment of homogeneity criteria (lot definition) 3. Establishment of acceptance criteria (in- spection plans, sampling plans) 3. Determination of methods of inspection (gaging, testing, visual inspection) 4. Classification of defects. The specification provisions for quality assur- ance must be formulated with care in order that maximum assurance of satisfactory quality may be obtained at the minimum cost that is consis- tent with the requirements of safety and effi- ciency of the end item. Incorrect classification of defects, unrealistic or ambiguous acceptance criteria, incomplete analysis of quality desired, and wrong methods of inspection may result in unreliable, costly, or hazardous ammunition, and render difficult the satisfactory fulfillment of a contract. 5-2. Amount of Inspection. The design engineer, unfamiliar with the practical aspects of inspec- tion, may reach the conclusion that in order to obtain materiel of satisfactory quality, accept- ance must be based on 100 percent inspection for every defect which is likely to occur. Ac- tually, this is not the case, unless it is essen- tial that there be no defective pieces accepted. Four factors militate against the performance of 100 percent acceptance inspection of Ord- nance materiel. a. The cost of 100 percent inspection (or "screening") of all materiel would be prohibi- tive, unless suitable automatic machines of proven reliability are available. b. Biecnuse of the extent of human fatigue associated with the inspection of large lots, 100 percent inspection by other than automatic ma- chines is seldom 100 percent effective. c. The contractor would tend to rely on the Ordnance inspection to screen out defectives, and would fall to inspect his product adequately; yet, inspection, is properly his own responsi- bility. d. When inspection testing is destructive, 100 percent inspection obviously is impossible. In most cases, adequate quality control may be obtained by a lot-by-lot sampling inspection; that is, a predetermined number of units of the product are selected from a lot in such a man- ner that the quality of the sample will represent as accurately as possible the quality of the lot. Normally, every effort should be made to select a sample consisting of units of product selected at random from the lot. 5-3. Definition of Lots. A lot is an aggregation of objects that are essentially of the same kind, size, form, and composition. A homogeneous lot is one in which the units of product are so thoroughly mixed that all por- tions of the lot are essentially alike. A lot may be homogeneous, but the units of which it is composed may not be identical. For example, a lot produced on an automatic machine may be homogeneous, but not all the units will be identical. On the other hand, the ingots poured from one heat of steel may be considered iden- tical in their chemical composition. 5-1
Homogeneity at lot and randomness of sample are closely related. If a lot of units is thor- oughly mixed, that is, homogeneous, each unit has an equal chance of being located in a cer- tain part of the lot. By randomness of sample is meant the selection of sample units in such a manner that each unit of the lot has an equal chance of being selected. It follows, therefore, that any sample drawn from a homogeneous lot is equivalent to a random sample drawn from a heterogeneous lot. The purpose of restricting lot size in the speci- fication is to assure homogeneity or uniformity by controlling the material which goes into the product, and the conditions under which the pro- duct is produced. However, if all variables which may make for some degree of nonuni- formity are strictly controlled, the resulting lot may be so small that the cost of inspection, subsequent handling, and use is disproportion- ately high. It follows, then, that only the major sources of variation, with respect to the specific product requirements, ought to be controlled. In practice, it is often difficult, if not impos- sible, to select perfectly random samples. For example, if a lot of 20,000 components is of- fered for inspection on trays containing 100 components in each tray, for practical reasons it may be necessary to treat each tray as a sub- lot, and select one sample unit drawn at ran- dom from each tray. Thus, the lack of complete assurance of homogeneity (a result of control- ling only major sources of variation) is counterbalanced by partial randomness of sampling. 5-4. Sampling Risks. In any form of accept- ance sampling, there is the inherent risk that a lot of acceptable quality will be rejected, while a lot of rejectable quality will be accepted. Fundamentally, there are three criteria by which a sampling plan should be judged to en- sure that the plan selected is the correct one to use. These criteria are: a. How the plan will operate with respect to lots of acceptable quality; b. How it will operate on lots which should be rejected; c. How it will affect the cost of inspection. The first two criteria may be determined by calculations involving the probabilities of ac- ceptance. In general, the risks of making a wrong decision, that is, accepting a bad lot or rejecting a good lot, may be reduced by increas- ing the sample size. However, in so doing, the costs of inspection are increased. 5-5. Operating Characteristic Curve. Four im- portant curves tell how any specific sampling plan operates with respect to lots that are de- tective in various percentages. Of these, the operating characteristic (ОС) curve is the most important because it gives an adequate picture of the plan's severity anddiscriminatory power. It pictures the probability of acceptance Pa for lots of various qualities, p' percent defective. The exact probability of finding d defectives Jn a sample of n pieces drawn from a lot of N pieces containing D defects is determined by the hypergeometric distribution, and is given by the formula _ N-Dcn-d Dcd О . a ------------- The first factor of the numerator is the number of ways in which (n — d) good pieces may be drawn from (N — D) good pieces in the lot, and the second factor is the number of ways in which d defectives may be drawn from D defectives in the lot. The denominator is the number of ways in which n pieces in the sample may be drawn from N pieces in the lot. The use of this hypergeometric formula is te- dious. For example, in a sampling plan where n * 150, and c (the acceptance number) - 4 defectives, applied to a lot of N 3,000 pieces of p' я 1 percent defective, gives a probability of acceptance p = 2970C150 + 2970C149 ' 30Cl + 3000C150 3000C150 297QC146 ' 30^4 (1) + 3000C150 Where n £ 0.1N, the binomial approximation to the hypergeometric distribution may be used. Thus, fcr the above plan ₽a - i50e0(°-M)150+ 15OC1(° 14в<0 01> * . . . + 150C4(0.99)146(0.01)4 (2) However, this is still a somewhat complicated calculation, unless a set of "Tables of the Bi- 5-2
nomial Probability Distribution" is available. Whenever n £ 0.1N and 2» 20 and p' 5 5 per- cent, use is made of the Poisson distribution, which may be used as a distribution in its own right, or as a good approximation of the bi- nomial. Using the Poisson distribution, the probability of acceptance is given by Pa s e’1*8 e’1,s1.5 + e"1B1.52/2 < e-l-W/e + e-b51.54/24 This is much easier to calculate. Also, tables of probabilities of accentance for various values of np' are to be found in any standard text on quality control. (The tables use p' as the frac- tion defective.) It should be noted that the hypergeometric and binomial probabilities are based on a fixed lot size, while use of the Poisson distribution as- sumes an infinite lot size. Since in most ap- plications in Ordnance the actual size of the lot is indeterminate at the time the sampling plan is developed, the Poisson is nearly always used. For further study, reference is made to "Sta- tistical Quality Control" by Grantor "Engineer- ing Statistics and Quality Control" by Burr. The ОС curve is a plot of percent defective ver- sus probability of acceptance for a given sample of n size and acceptance number of c defectives. Thus, for any given percent defective (normal- ly the abscissa), the probability that a lot of this quality will be accepted may be found. Fig- ure 5-1 is a plot of a typical ОС curve. Terms frequently used to indicate the charac- teristics of a plan, or used as an index to a series of plans, can best be indicated by ref- erence to the ОС curve. The term used most often is Acceptance Quality Level (AQL). The AQL represents that quality, expressed in terms of percent defective, which is considered ac- ceptable and which will be accepted most of the time. The probability of not accepting material of this level of quality is called the producer's risk (a). In standard practice it is customary to base the AQL on a probability oi acceptance of 95 percent, with a producer's risk of rejec- tion of 5 percent. However, these figures are Figure 5-1. Typical Operating Characteristic (ОС) curve purely arbitrary. Plans outlined in MIL-STD- 105A are based on acceptance of AQL product of quality from 88 percent to 99 percent of the time, with corresponding producer's risks of 12 percent and 1 percent. On the other hand, the Lot Tolerance Percent Defective (LTPD) is the quality, expressed in terms of percent defective, that is considered unacceptable and which will be rejected most of the time. The consumer's risk {p) is the prob- ability of accepting material LTPD quality. In practice, p usually is given a value of 0.1, or 10 percent, but other values may be given if considered desirable. The discriminatory power of a plan cannot be determined merely by knowing the AQL. The proximity of the AQL and LTPD of a plan in- dicates its ability to distinguish between good and bad quality. This is reflected by a steep ОС curve. A tight AQL may not offer sufficient protection if the LTPD is poor. 5-6. The Average Outgoing Quality (AOQ) is the average quality of a succession of lots which have been accepted. U the rejected lots are not resubmitted for inspection, then, disre- garding the removal of defectives found in the samples, the average quality of the lots ac- cepted is the same as the quality of the lots 5-3
submitted. For example, if 1,000 lots of equal- ity 6 percent defective are submitted, then under the sampling plan Indicated by the ОС curve in figure 5-1, 620 lots will be accepted; their av- erage quality will be в percent defective. If from the ОС curve a curve of AOQ values is plotted, it will be found to be a straight line, for the average outgoing (accepted) quality is always equal to the average incoming (submitted) qual- ity. This is true in the case of destructive tout- ing when rejected lots are scrapped. However, when rejected lots are screened 100 percent and resubmitted for inspection, the AOQ of the accepted lots will be better than the av- erage incoming quality. For example, if 1,000 lots of a quality в percent defective are inspect- ed under the plan indicated in figure 5-1, 620 lets will be accepted and 380 lots rejected. Af- ter the rejected lots are screened and finally accepted, the quality of these lots will beO per- cent defective. The AOQ for the total lots will then be 020 * ° * ° = 3 72 Percent 1000 which is better than the incoming (submitted) quality. If from the ОС curve (figure 5-1) the AOQ val- ues for the range of qualities are computed and plotted, the AOQ curve will be similar to that shown in figure 5-2. it will be noted that there is a maximum value known as the Average Out- going Quality Limit (AOQL), which is the («ecewT DEFECTIVE) Figure 5-2. Typical сип е of Average Outgoing Quality (AOQ) poorest average quality that the plan will ac- cept. Obviously, the quality of the accepted or outgoing lots can never be worse than that of the incoming or submitted lots. 5 -7. Establishing the Acceptable Quality Level (AQL). The ideal goal of procurement is to ac- cept only perfect materiel. However, several factors involved in the attainment of such per- fection make it necessary that something less than perfection be accepted. First of all, if a manufacturer were forced to submit perfect materiel at all times, his manufacturing costs would be very high, and the resultant cost would be high. Secondly, to be assured of buying only perfect materiel, 100 percent inspection on all lots submitted for acceptance would have to be made. This would make the cost of inspection prohibitive. To maintain manufacturing expense and cost of inspection at a reasonable level, less than perfect materiel must be accepted. The question then arises: kt what point will the combined cost of inspection and manufacture be reasonable, while still primarily assuring the acceptance of good materiel? This question should be answered whenever an AQL is set. The AQL may not be the perfect answer in all cases, however, because of the numerous fac- tors which must be taken into account when es- tablishing an AQL. In establishing an AQL the most important con- sideration is the seriousness of the defect. The degree of compromise made with respect to the quality considered acceptable is completely de- pendent upon this factor. Systems of classify- ing defects assist in permitting defects of simi- lar natures to be treated alike. One of the factors to be considered when estab- lishing an AQL is the degree of manufacturing difficulty associated with the item. Cost is in- directly involved in this consideration, as costs increase with greater manufacturing difficulty. To determine the degree of manufacturing diffi- culty for a particular item, any knowledge of actual expe* nee with the item should be put to use. For purposes of setting the AQL in those cases where no data exists, the degree of manufacturing difficulty is associated with the number of defects listed under one classifica- tion. It is obviously more difficult to control five characteristics than one, thereforethe AQL would be less stringent for five than for one. Assuming the defects are independent of each 5-4
other, as the number of defects listed under one classification increases, the probability that all of them will occur with the same frequency de- creases. When setting an AQL for a component part it is important to view the item as a whole, in order that the importance of different features can be seen in proper perspective. For example, con- sider the case of a squib that is to be a com- ponent of a rocket. The AQL for the functioning test of the rocket is determined to be 2.5 per- cent. This means that the AQL must be more stringent for the functioning test of the igniter (1.5 percent) and still more stringent for the functioning test of the squib (0.65 percent). This must be done in order to avoid penalizing the manufacturer of the igniter (the functioning of which depends on the squib) and the manufac- turer of the rocket (the functioning of which de- pends on the squib and the igniter). However, if xt is known that the same squib is to be used in a J A TO unit, for which the functioning test has an AQL of 0.25 percent, the AQL for the squib in this case should be about 0.10 percent. Obviously, the same squib cannot have two AQL's; therefore, the squib intended for the rocket also must have an AQL of 0.10 percent, unless a method of grading squibs can be de- vised. In all cases the component snould be given the most stringent AQL required for any use. 5-8. Classification of Defects. Defects of simi- lar importance are treated alike. To accom- plish this it is necessary to classify them into groups, the number of which is a compromise between the degree of selectivity desired and the administrative complexity. The following excerpts taken from MIL-STD-105 reflect the prevalent classes and definitions for the classes. a. Method of Classifying Defects. A classi- fication of defects is the enumeration of pos- sible defects of the unit of product classifie ’ according to their importance. A defect is ai deviation of the unit of product from require- ments of the specifications, drawings, purchase descriptions, and any changes thereto in the contract or order. Defects are normally grouped into one or more of the following classes; however, the Government reserves the right to group defects into other classes. b. Critical Defects. A critical defect is one that judgment and experience indicate could result in hazardous or unsafe conditions for in- dividuals using or maintaining the product; or, for major end-item units .uf product, such as ships, aircraft, or tanks, a defect that could prevent performance of their tactical function. c. Major Defects. A major defect is a de- fect, other than critical, thst could result in failure, or materially reduce the usability of the unit of product for its intended purpose. d. Minor Defects. A minor defect is one that does not materially reduce the usability of the unit of product for its intended purpose, or is a departure from established standards having no significant bearing on the effective use or op- eration of the unit. 5-9. Inspection by Attributes is the grading of a unit of product as defective or nondefective, with respect to a given requirement, by sampling in- spection. When attributes inspection is per- formed, the decision to accept or reject is based on the number of sample units found defective. The extent of deviation from the requirement is not considered in.this form of inspection. This form is simple to administrate, normally is simple and rapid to perform, is not dependent upon a particular distribution of product to pro- vide a desired assurance, and involves no math- ematics to determine lot acceptability. It is al- ways possible to use attributes inspection, and in many instances, such as in GO and NOT GO gaging, it is the only type possible. 5-10. Single-Sampling is a technique in which only one sample of n items isinspectedto reach a decision on the disposition of the lot. A sam- pling plan is best described by the sample size (n) and rcceptance number (c). When the num- ber of defectives found in this sample equals, or is less than, the acceptance number pre- scribed by the sampling plan, the lot is accepted. U the number of defectives exceeds the accept- ance number, the lot is rejected. Anexampleof a single plan is as follows. Select a sample of 100 units. If 3 or fewer defectives are found, accept the lot; if more than 3 are found, re- ject the lot. Single plans may be summarized as follows: Sample size Acceptance no. Rejection no. 100 3 4 5-11. Double-Sampling is a technique in which a second sample of n items is inspected when
Um results of the first sampling do not accept or reject the lot. Depending on what is found in the first sample, there are three possible courses of action. The lot may be accepted, it may be rejected, or the decision may be defer- red until the results of a second sampling are obtained. A decision is sure to be made on the basis of a second sample, inasmuch as the re- jection number of a second sample is one more than the acceptance number. An example of a double plan is as follows. A sample of 75 units is selected. If the sample Contains 4 or fewer defectives, the lot is ac- cepted. If the sample contains 9 or more de- fectives, the lot is rejected. If more than 4 defectives, but fewer than 9, are found, a second sample of 150 units is selected and inspected. If In the combined sample of 255 units fewer than 9 defectives are found, the lot is accepted; but if 9 or more defectives are found, the lot is rejected. Double-sampling plans are sum- marized as follows: Cumulative sample Acceptance no. Rejection no. 75 4 9 225 8 9 5-12. Multiple-Sampling is a procedure in which a final decision to accept or reject the specific lot need not be riade after one or two samples, but ma^ require the drawing of sev- eral samples. An example of a multiple-sam- pling plan is as follows: Cumulative sample Acceptance no. Rejection no. 30 2 8 60 8 13 90 12 18 120 17 22 150 21 27 180 27 32 210 35 36 The inspector selects a sample of 30 from a lot of 4,000 items. If the first sample contains 2 or fewer defectives, the lot is accepted; if 8 or more defectives appear in the sample, the lot is rejected. If more than 2 but fewer than 8 defectives appear in the sample, the inspector draws a second sample of 30 items. If, in the total sample of 60 units, 8 or fewer defectives are found, the lot is accepted. If 13 or more defectives are found, then the lot is rejected. If more than 8 but fewer than 13 defectives are found, a third sample of 30 is inspected. This process Is continued, until a decision is reach- ed, which in this case will be not later than the seventh sample. It is a common misconception that a compar- able double- or multiple-sampling plan is less stringent than a single plan. This is not true, since the acceptance and rejection numbers are selected to make them equivalent. The ОС curves for equivalent single, double, and mul- tiple plans are shown in figure 5-3. 5-13. ОС Curves for Comparable Single-, Double-, and Multiple-Sampling Plans. It will be noted that the ОС curves of the three plans almost coincide, showing little difference in their severity and discriminatory power. How- ever, there is a great difference among the plans in the average number of pieces per lot that will be inspected before a decision is reached. Figure 5-4 shows the ASN (Average Sample Number) curves for the same plans. In the single-sampling plan the sample is al- ways completely inspected, regardless of a possible earlier decision. The ASN curve for the single-sampling plan is, therefore, a straight line. In the double-sampling plan the first sample is completely inspected, ^ut inspection of the second sample ceases as soon as the number of defectives found in the combined samples equals the rejection number. Since lots of very good quality will be accepted, and lots of very bad qu ility will be rejected on the first sample, and inspection of the second sample is curtailed, the average amount of inspection will be less for a double-sampling plan than for a single-sampling plan. As will be seen, mul- tiple-sampling reduces still further the amount of inspection. 5-14. Relationship of Sample Size to Lot Size. The sample size is of considerably more sig- nificance, with respect to severity and dis- criminating power, than the lot size from which it is taken. Figure 5-5 is composed of ОС curves, all having the same sample size and acceptance numbers, but different lot sizes. 5-6
Figure 5-3. Operating Characteristic (ОС) curves for single-, double-, and multiple-sampling plans It is readily observed that the ОС curves for lot sizes of 1,000 and over aro virtually equivalent to that for a lot size of infinity. Thus it is often said that the lot size has almost no effect on the sampling plan. This statement holds, pro- vided the sample size is a small fraction of the lot size (normally less than 10 percent). Fig- ures 5-6 and 5-7 demonstrate the effect of varying sample size for finite and infinite lot aize, respectively. Although percent sampling appears logicalfrom a cost or time aspect, it is decidedly not so with respect to the degree of protection offered. The absolute size of the sample, it has been noted, affects the characteristics of a plan more sig- nificantly than the lot size. A proportional change in both the sample and lot size, there- fore, will result in different degrees ot pro- tection. Figure 5-8 is an illustration of this phenomenon. b-7
100 SINGLE SAMPLING ou V» Ulb- OF PIECI TIOM LO" » > DOUBLE^ SAMPLING BE NUMBER PER INSPEG h « > < “T1 MULTIPLE SAMPLING ASN «AVER Al INSPECTED » 4 5 < «V 0 c > 2 3 d 1 6 p'l PERCENT DEFECTIVE Figure 5-4. Average Sample Number (ASN) curves for plans shown in figure 5-3 The effect of varying the sample size can be observed by noting figures 5-6 and 5-7. Ob- viously the greater the sample size, the more stringent and more discriminating will the plan be for a fixed acceptance number. Similarly, for a fixed sample size, the severity of the plan increases as the acceptance number decreases (figure 3-9). 5-15. Acceptable Quality Level as Basis for Ingpectign. Usually, the effect of a sampling plan for acceptance Inspection is to force man- ufacturers to supply materiel of such good qual- ity that only a very small proportion of the lots is rejected. The manufacturer, when given the AQL, knows the major piece of inf or mat ion per- taining to the sampling plans, which will affect the disposition of his lots. The manufacturer will know that by submitting products as good as, or better than, the AQL, rejections can be kept to an absolute minimum. Й the manufac- turer produces worse than AQL level, many inspection lots will be rejected, with increased costs necessary to screen or to scrap the lots. On the other hand, a production quality level better than the AQL may involve higher pro- duction costs than are necessary. The know- ledge of the AQL level will enable the producer to manufacture most efficiently. If some point other than the AQL is set as a basis for inspection, the acceptable quality level will differ among manufacturers, depending on the particular sampling plan used. 5-9
Figure 5-5. Effect of lot size on sampling plan Figure 5-6. Effect of sample site with fixed acceptance number and finite lot size Figure 5-7. Effect of sample size with fixed acceptance number and infinite lot size Figure 5 -8. Fallacy of sample size based on fixed ratio to lot size 8-9
Certain situations may arise where high assur- ances of achieving specific reliabilities are de- sired. In such cases, the utilization of some level, other than the AQL, would be necessary to achieve such desired protection. Thus if 85 percent assurance of achlev ' a reliability of BB percent were denired, it wo .id be similar to a probability of 5 percent at accepting lots 1 percent or more defective. This point would then be the basis for inspection. 5-16. Resubmlsslpn and Retest. The distinc- tion between resubmlnsion and retest is fre- quently lost in applying sampling procedures. It is true that both resubmission and retest re- quire additional inspection, but the similarity ends there. A retest is essentially a double - (or multiple-) sampling procedure. It acknow- ledges the fact that the first sample is not large enough to distinguish adequately between ac- ceptable and nonacceptable materiel in border- line cases. When materiel is subjected to a retest, it should not have been altered in any way prior to retest. The materiel subjected to retest has not yet been rejected or accepted. The risks involved in the case of retest are those of a specific plan and are, therefore, clearly defined. Resubmlssion, on the other hand, is a proce- dure for reinspccting materiel that has pre- viously been rejected. It implies (and the pro- cedures for resubmlssion should require) some procedure for removing the defective portions of the lot or reworking the lot. To prevent the resubmission of nonscreened or nonreworked materiel, the acceptance criteria for n resub- mitted lot should be more stringent than those applied when the lot was originally submitted. 5-17. Continuous-Sampling Plans. It has be- come evident that where production is contin- uous, as in conveyor lines, the formation of in- spection lots for lot-by-lOt acceptance is some- what artificial, and may be impracticable or un- duly costly. Some other plan of inspection, therefore, is necessary for this type of produc- tion. Continuous-sampling has been developed to fill this need. Continuous-sampling dispenses with the notion of the lot as a static entity, and con- siders instead a continuous flow of the product. The essence of most continuous-sampling plans is the qualifying of the product for sampling Figure 5-9. Effect of acceptance number with fixed sample site inspection by inspecting a large number of con- secutive units of the product as produced, and then the sampling of the product that is quali- fied. One of the major virtues of many of these plans is that products which have passed the inspec- tion station are accepted, and are free of re- inspection. This type of plan attempts to as- sure that the average outgoing quality will not exceed a specified limit where the process is in control. It is recognized that in some- continuous-sam- pling plans, questions arise as to the quality of the product between tne time that a defect is detected and the last preceeding sample. How- ever, one of the assumptions of many of these plans is that in the long run, the quality will be no worse than quality specified by the plan. Such inspection schemes are advantageous to the manufacturer because he does not risk hav- ing large quantities rejected, does not need storage facilities for static lots or rejected ma- teriel, and does not have to tie up his production lines tor reinspection a .d reprocessing. Such plans also work to the consumer's advantage, because they result in lower cost. 5-10
There are certain disadvantages to the con- sumer associated with continuous-sampling. One is that many of these sampling schemes do not impose a penalty upon the manufacturer for poor quality. Since no penalty is imposed, there is little or no incentive to precontrol the qual- ity of product. The only added burden in this respect is to increase the number and/or length of qualifying periods. A second disadvantage is that unless the frequency of sampling is more than 2 percent of the production, there is little or no chance of detecting spotty quality. This is particularly important where significant de- fects are concerned. 5-18. Statistical Aspect of Parallel Design. Parallel design is a term applied to a situation where more than one route or train is provided from initiation to result. Initiation asamsaResult Parallel design (three trains) Because in conventional design, functioning of the item is dependent on only one train from the point of initiation to the result, the overall re- liability of the item is limited by the functional quality of each of the various elements com- prising the train. These limitations, concerned with quality of functioning, are occasioned by failure to achieve perfection in the design, manufacture, and inspection of the component. Since perfection is unobtainable, and costly to approach, there is an obvious limitation on the extent to which reliability of the end item can be improved through increasing the quality of the components. In cases where exceptionally high assurance of functioning or safety is de- sired, a parallel design may be the only eco- nomical way to achieve these. If we have an item with two trains in parallel, the probability of the item's not functioning is squared, as com- pared with the probability of not functioning for a conventional design. With three trains in parallel, the probability is cubed, and so on. Thus, a parallel design permits considerable saving in cost of production and inspection. In cases where this saving outweighs the cost of duplicating certain aspects of the design, the use of a parallel design ought to bo considered. The following are examples which illustrate the advantages. a. If in a lot of 1,000 conventionally designed items we desire an assurance of 0.09 of reject- ing lots with a fraction defective greater than 0.002, it is necessary to inspect a sample of 900. If two trains in parallel are used, it is possible to maintain the same level by inspect- ing a sample of 140. b. If we desire a fraction defective less than 0.0001 in the end item, it may be obtained by the use of two trains in parallel, each train hav- ing a fraction defective of less than 0.010. 5-19. Relationship Between Sampling Plan, Tolerance Limits, and Safety Factor. (See fig- ure 5-10.) In the caseof primary requirements, which define the physical shape or composition of an item, the sampling procedures can, if nec- essary, be established to require a level of quality higher than that being currently attained in production. The decision in this case is an economic one, as setting of such standards may force the producer to screen his product, or to develop a different method of production. One must determine whether the cost of this action is overbalanced by the value of the result ob- tained. However, in the case of a secondary requirement the results obtained measure not only the effectiveness of the producer, but also the effectiveness of the designer. Most func- tional requirements have this property, and considerable care should be taken when deter- mining both the design and acceptance para- meters that the producer is not required to meet functional requirements which have little correlation with the primary requirements, or vice versa. The best approach to thia problem is to obtain data from the functioning test, ana- lyze the data statistically, develop tolerance Figure 5 -10. Relationship of eatnpling plan, tolerance linrite, and egfety factor 5-11
limit* within which lie a certain percentage ct the distribution, determine whether such limits are suitable from a functioning standpoint, and, if so, develop acceptance sampling procedures baaed on the same distribution percentages and risks as were used to estimate the tolerance limits. И a safety factor is incorporated in the speci- fied limits, the test becomes one of increased severity, and the point of test is not the real point of interest. The problem of developing the acceptance requirement in this case is that of determining the reliability and confidence needed at the specified limit in order to guaran- tee the reliability and confidence desired at the point of interest (specified limit minus safety factor). Data is necessary for resolution of thia problem, for it is necessary to know the distance between the specified limit and the limit of interest, in terms of certain para- meters of the estimated distribution. 5-20. Sampling Plans Based on Variables. In- spection by variables simply requires each item in the sample to be measured and the reading recorded. Thue, the exact also c* character- istic is Imown, and how much variation occurs from sample unit to sample unit is also known. It can also be determined how much each meas- urement varies from the specified limits. After the readings are obtained, measures of the av- erage and dispersion are computed. To deter- mine the acceptability of each lot, the average and. a multiple of the measure of dispersion are > compared with the specified design limits. Since ; most production processes are normal distri- bution, or nearly so, one can determine fairly accurately what percentage of the units in the lot will exceed any given limit. 5-21. Variables Inspection Compared with At- tribute Inspection. In attribute inspection each item is measured^ and classified as either good or bad. Often this measure may be the GO, NOT GO type of check. No consideration is given to items just outside the limits, or to the extent of variation from item to item. Thus, one advan- tage of this method is the ease and simplicity of checking each attribute. Little skill is nec- essary, and no calculations are required. However, sampling by variables requires the actual measurement of each item, and the mathematical computation of measures of cen- tral tendency and of dispersion. The most com- monly used measures of central tendency and of dispersion are the mean (arithmetical average) and the standard deviation. This gives valuable information about the items under consideration. It is not necessary to inspect as many items under the variables plan in order to obtain the same assurance of accepting only good lots. If the item requires a great deal pf labor to in- spect it, or if it is damaged during inspection, the variables plan would be preferred. Certain limitations exist for variables inspec- tion. The distribution must be normal or near- ly normal. If it Is abnormal, the actual char- acteristics of the sampling plan will differ from those upon which the plan was based, and may result in unnecessary acceptance of materiel of inferior quality or rejection of an acceptable product. If screening has previously been per- formed, some lots of acceptable quality will be rejected by the variables inspection plan. The reason for this is that in screening a small fraction of the distribution has been removed, but the procedure for determining acceptance is based on the assumption of normality and can- not recognize the effect of the screening opera- tion. REFERENCES AND BIBLIOGRAPHY 1. Grant, "Statistical Quality Control." 2. Burr, "engineering Statistics and Quality Control." 5-12
EFFECT OF DIMENSIONING AND TOLERANCING ON INSPECTION 5*22. Introduction. When an Ordnance item leaves the design and engineering phases and goes into production, dimensional control is exercised by inspectors who determine, by gag- ing and, when necessary, measurement, that the items and components conform dimensionally to the requirements specified by the design en- gineer. The dimensioning and tolerancing of an item impose difficulties on production and in- spection, unless presented in a proper manner. Department of Ordnance Drawing 30-1-7, "Standard for Dimensioning and Tolerancing," has been established as a standard for dimen- sioning and tolerancing of ammunition items, and the design engineer should understand thor- oughly the methods outlined therein and their practical application, in order that difficulties may be avoided in the production and inspection stages. The application of an incorrect symbol or requirement may result in the manufacture of a component that is not what the designer in- tended, but which the inspector must accept, since it complies with the drawing. The definition a dimensioning terms is clearly stated on page 2 of Drawing 30-1-7. The en- gineer should bear in mind that a basic dimen- sion, although exercising control, is not checked directly by the inspector. A reference dimen- sion, being informative only, is not a control dimension, and is not checked by the inspector. Generally speaking, only toleranced and datum dimensions are checked by the inspector. 5-23. Locational Tolerance Symbols. There are two classes of locational tolerance symbols permitted, independent and dependent, and the difference between them is of great importance. 5-24. An Independent Locational Tolerance is a fixed tolerance to whi^h the manufacturer must adhere, whether the part produced is of maxi- mum permitted size or minimum permitted size. (See paragraph 5-25, on dependent loca- tional tolerances, for comparison.) Independent locational tolerance requirements are specifi- cally checked by the Inspector independently of other requirements, normally by the use of dial indicating gages. The application of the various symbols, and the implications of their use, should be thoroughly understood. The use of the datum surface symbol I- P -I in conjunction with the tolerancing symbols, must be carefully studied to assure that what is speci- fied on the drawing is in fact the requirement which is needed. Examples of the requirements and use of the symbols follow. The concentricity symbol |6Pxxx| indicates that the surface to which it is applied, if made per- fectly, will have a common centerline with the datum surface. Concentricity must not be con- fused with ovality (out-of-roundness), which ip controlled by the diametral tolerance on each surface. Example; the drawing of the illus- trative piece shown on page 7 of Drawing 30-1-7 is reproduced here in figure 5-11. The two con- ditions which may prevail (ovality and eccen- tricity) are shown in figure 5-12. The illustra- tion in figure 5-12a shows concentricity, but maximum permitted ovality; while figure 5-12b shows maximum permitted eccentricity, but perfect roundness. OR 004 1 Figure 5-11. Concentricity requirement If ovality is inherent in the part (thin-walled sections of large diameters), the tolerance for concentricity should not be less than the sum of the diametral tolerances, unless a requirement for maximum ovality is included, because the inspection will not discriminate between ovality and eccentricity ualess laboratory methods are used. If the concentricity tolerance is less than the sum of the diametral tolerances, the inspec - tor will reject the parts for falling to conform to the concentricity requirement (the diametral gages having accepted the parts), in spite of the fact that the parts might be perfectly concentric but out-of-round. 5-13
A. OVALITY Figure 5-12. Ovality and eccentricity tolerances In general, the concentricity symbol should be used only when two or more surfaces are Involved. When a single surface is involved, the desired conditions should not be stated in terms of concentricity, but of some other attribute, such as straightness. The symbol for perpendicularity | i Axxx| indi- cates that the surface or form to which the sym- bol is applied must be perpendicular to the datum surface indicated by Г-'А. One ex- ample is shown on page 8 of Drawing 30-1-7. Another example is illustrated here in figure 6-13. Figure 5-13a shows the intended conditions in a cone, that is, the entire form is perpendicular to the base of the cone. Figure 5-13b shows the condition which is to be controlled and the method of checking. It should be noted that the cone in figure 5-13b is concentric but not per- pendicular. The condition shown cannot be con- trolled by concentricity requirements. However, concentricity requirements may be required in order to control the condition shown in figure 5-14a. This cone is perpendicular to the base, but not concentric with the diametral surface A. Figure 5-14b illustrates the correct toleranclng of a flared cone to obtain the desired results. Two examples of the use of the symbol for paral- lelism I И Axxx I are given on page 9 of Drawing 30-1-7, the first of which requires parallelism between two surfaces, and the second of which requires parallelism of a tapered form with a surface. However there are certain conditions of which the engineer must be aware. Figure 5-15a shows the conditions which the engineer desires, while figure 5-15b shows a condition which meets the requirement for parallelism, but which may not be desired. The normal inspection process will not dis- criminate between lack of parallelism and the condition shown in figure 5-15b. Rejection would be on the basis of nonconformance with the requirement expressed in figure 5-15a, whereas in fact the piece shown in figure 5-15a does meet the requirement*. In the case of a very light part exhibiting the condition shown in figure 5-15b, the inspection process might accept the piece if the convex surface is resting on the surface plate, because the spring pressure of the dial indicator might cause the piece to roll in the direction that the indicator is moved. Consideration should be given to a requirement for flatness of the datum surface, in addition to the requirement for parallelism. The symbol t indicates either symmetry or centrality. Symmetry is the equal distribution of a form about a centerline. Centrality is the spacing of surfaces equidistant irom a center- line. In figure 13 of Drawing 30-1-7, the tongue at the bottom of the piece is symmetrical about the centerline of the datum surface I -5,41. The slot at the top of the piece is equally distributed, that ie, the sides of the slot are equidistant from the centerline of thedatum surface • In actual practice, there are few occasions in 5-14
the design of ammunition items when the inde- pendent symbol I is applicable. However, in many instances where assembly is the prime requisite, a dependent locational symbol indi- cating a requirement for centrality may be used to advantage, and a discussion of this point is to be found in paragraph 5-25. It should be emphasised that the independent symbol itP.OOil refers to total variation and not to displacement of centerlines. This is easily understood if it is remembered that the centerline of the designated form may vary 0.002 on either side of the centerline of the da- tum form, thus allowing a total variation of 0.004. READING NOT TO EXCEED .004 WHEN ROTATED 380' Figure 16 of Drawing 30-1-7 illustrates a joint requirement for concentricity and perpendicu- larity. The prime requirement is for concen- tricity of the 0.874 diameter with the pitch di- ameter specified as the datum I- F -I. The secondary requirement is that the effect of lack of perpendicularity between the 0.874 diameter and the end surface ! - Й -П shall be included with the eccentricity in the total variation of 0.004 which may be indicated on the dial. Since there is no specific requirement for per- pendicularity^ surface I- R -] with the pitch diameter I- P - lf the combined requirement in- dicates that if the surfaces are concentric, then the play of the mating threads may allow sur- face Г- R A to be out of perpendicular with the pitch dinmeter within 0.004 on the dial reading taken at the opposite end of the piece. Figure 5-18 shows schematically the effect of each re- quirement, and of the combined requirement. Figure 17 of Drawing 30-1-7 shows the prime requirement of concentricity of the threaded counterbore with the datum counterbore, and the secondary requirement of perpendicularity of the counterbores with the end surfaces - I and I - 8 -I. In addition, there is an implied requirement that the effect of lack of parallelism of the axis of the screw thread with the center- line of the base counterbore shell be included in the total variation. This is implied by the sym- bol , which indicates that the total vari- ation must not exceed 0.004 when readings are taken at gaging positions located 0.25 and 2.50 inches beyond end surface Г- 8 Я. The greater length usually compares with the length ol the mating part. If neither dial exceeds 0.004, then 8 Figure 5-i3. Requirement for perpendicularity the part is acceptable. Figure 5-17 shows the effect of cocked centerlines on dial readings. 5-25. A Dependent Locational Tolerance is a variable, since the effect of associated dimen- sional tolerances affects the magnitude of the locational tolerance. Dependent locational tol- erances are inspected by functionalgagee, which simulate the mating parts. This type of tol- erance should be used when facility of assem- bly is the prime factor. Examples of dependent tolerance symbols are shown in Drawing 30-1-7, pages 12 through 16. 5-15
Figure 5-14. Eccentricity of cone with base and correct method of dimensioning Figure 5-15. Two conditions of parallelism Figure 5-16 Effects of concentricity and perpendicularity 5-16
Figure 5-17. Effect of cocked center lines on dial readings Dependent locational tolerances should be used, rather than toleranced coordinates, whendimen- sioning the location of holes. Figure 23A of Drawing 30-1-7 explains why this is desirable, but further explanation is given in the following paragraphs, since production and inspection are vitally interested in the method used. Figure 5-18a shows a portion of simple part with a single hole, the location of which is fixed by coordinates. Figure 5-18b is an enlarged view of 5-18a. The square in the center shows the permissible tolerance zone in which the center of the hole may vary. The large circles show the positions of the minimum holes at the positions of maximum variation permitted by coordinate dimensions. In order to permit the manufacturer his full tolerance, the gaging pin would necessarily be of the shape shown by the four arcs that are drawn in heavier lines. Since the cost of making simple locational gages with pins of that shape would be prohibitive, the gage engineer will design the gage with a round pin. To ensure that only conforming parts will be accepted, the pin would necessarily be of the size shown by the dashed circle. The Figure 5-18. Effect of hole location of coordinates on production and inspection 5-17
corresponding tolerance zone permitted by this gage is shown by the circle inscribed in the square, and the dark shaded areas of the square indicate the extent of the permissible tolerance zone that will be denied the manufacturer. On the other hand, a gage to permit the manu- facturer the full tolerance expressed on the drawing would necessarily contain a round pin of th* iize indicated by the dotted circle. The cerrebnonding tolerance zone will be that circle in which the square is inscribed, and the light shaded areas indicate the extent to which parts xit of tolerance limits will be accepted by the gage. * The same part, dimensioned with a dependent locational tolerance, is shown in figure 5-19a, and the explanation of it in figure 5-19b. The intersection of the basic coordinates, the dimensions of which are theoretically exact, locates the theoretically exact center of the hole. Since nothing can be produced to exact dimen- sions, the small circle represents the locational tolerance area 0.002 in diameter, within which the actual center of the hole may vary when produced to minimum size. Since there are an infinite number of circles, representing the minimum hole, with centers located on the tol- erance circle, representing maximum permis- sible locational variation, the pin of the gage will be made to a size equal to the minimum dia- meter of the hole*minus the locational tolerance I 0.002 ). The difficulties encountered when tol- eranced coordinates were used have been over- come; the manufacturer has his full tolerance, and the gage will accept only conforming parts. Figure 5-20 gives a graphic explanation of why the gage will accept the maximum hole even when the tolerance circle is increased by the amount of the tolerance on the diair eter of the hole. In the sketches, shown enlarged in the figure, the circular shaded area is the pin. The solid and dashed circles represent the extreme posi- tions of the hole in which the gage pin will eiv ter, the corresnonding centerlines being shown figure l>-19 Dependent locational tolerance 5-18
Figure 5-20. Gaging pin in maximum and minimum holes solid and dotted. The areas In which the centers of the maximum and minimum holes may vary are shown as circles inscribed in the center at the pin. The diameter of the tolerance zone in which the minimum hole may vary is 0.002, shown on drawing as |0.QQ21. The diameter of the tolerance zone in which center of the maxi- mum hole may vary is 0.005, the locational tol- erance 0.002 plus the diametral tolerance 0.003. The design engineer should bear in mind that the gage engineer will employ the dependent locational tolerance system, since he will specify round pins in the gages, regardless of whether the location of the holes is shown by tol- erance coordinates, except in special applica- tions. In order to explain the definition cd a de- pendent locational tolerance symbol when shown as I P.0041, figures 20 and 20A of Drawing 30- 1-7 are reproduced here in figure 5-21. The geometry of the part determines the re- quirement, and since the part is cylindrical it is obvious that the requirement IP.0d4l refers Г ₽J004 A -------- Figure 5 -21. Part and receiver tn gaging requirements 5-19
to concentricity. The small diameter of the re- ceiver is 1.004, permitting the centerline of the small diameter of the part to be displaced 0.002 in any direction from the centerline of the large diameter, when maximum metal conditions pre- vail In the part. When minimum metal condi- tions prevail, the effects of the diametral tol- erance enter, and the centerline of the small diameter may be displaced 0.004 In any direction from the centerline of the large diameter. Fig- ure 5-22 shows the variations possible under minimum and maximum metal conditions. It should be noted that the effect of ovality is checked in the gaging operation. The methods of dimensioning and locating holes shown la figures 22, 23, and 24 of Drawing 30-1- 7 will easily be understood when the foregoing principles of dependent locational tolerancing are grasped. It should be borne in mind, how- ever, that not only will the positions of the holes in respect to the datum hole vary with the size, but they will also vary with respect to each other. A variation of the dependent locational tol- erance, involving basic angular and radial di- mensioning, is shown in figure 5-23a. In the part illustrated, the large hols is located by de- pendent locational tolerance in the manner pre- viously described. The small hole is located by a combination of a basic angular dimension and a basic radial dimension. Any attempt to use a toleranced angle and toleranced radius will re- sult in a tolerance zone similar to that shown in figure 5-23b. Since the gage engineer will use round pins, the dark shaded areas repre- sent the extent to which the contractor will be denied the use of the full tolerance when the gage accepts only conforming parts, while the light shaded areas represent the extent to which nonconforming parts will be accepted, if the figure 5-22. Receiver and fixed dimensions — effect of dimensional and locational tolerances 5-20
Figure 5-23. Application of basic angular and radial dimensioning contractor receives the full extent of the tol- erance zone. The use of the dependent locational tolerance will result in round tolerance zones benefiting both the producer and the gage de- signer. Figure 26 of Drawing 30-1-7 shows a dependent locational tolerance applied to the same part shown in figure 16 (of the same drawing) with an independent tolerance. It will be noted that with the independent tolerance the total variation was 0,004. With the dependent tolerance, the func- tional tolerance increases from 0.004 for maxi- mum metal conditions to 0.006 with minimum metal conditions. If the part is produced to minimum permissible size, the clearance be- tween the part and receiver will be greater. This condition will permit greater eccentricity and/or cocking effect due to lack of perpendi- cularity between surface R and the thread when surface R makes up against the slider plug of the gage. Figure 5-24 shows the eccentricity and cocking permissible when metal conditions are maximum, while C and D show the same conditions when metal conditions are minimum. The dependent locational tolerance should not be applied to a part without consideration of the implications. Only the effect of lack of per- pendicularity is included, both in the dependent and the independent tolerance applications to this part. In both cases the lack of perpen- dicularity is the amount of cocking permitted as a result of play of the mating threads. A separate perpendicularity requirement should be used when the degree of contact between seating surfaces is important. In figure 26 (Drawing 30-1-7) there is no requirement that surface R ehall seat 360 degrees on the gage plug. It might only seat 5 degrees, and if the threads were a close fit, the gage might still accept the part. Figure 5-25, which is an en- larged view of the area circled in figure 5-24d, explains why the gage may accept a part when the lack of perpendicularity is greater than the apparent effect. The dependent symbol shown in figure 26 of Drswing 30-1-7 is particularly applicable to mating parts that after assembly must fit into a third component or into a chamber. It should 5-21
A. MAX. ECCENTRICITY [ LZZZZZI •87в B. MAX. C0CKIN6 EFFECT MAXIMUM METAL CONDITIONS Г-.872 .003-^ I T FZZZZ OOS T\ZZZZZ C. MAX. ECCENTRICITY 0. MAX. C0CKIN6 EFFECT MINIMUM METAL CONDITIONS Figure 5-24. Effect of variation in site on locational tolerance not be used where good seating Is required, such as in the case shown below infigure 5-26a. The requirement should be shown as in figure 5-26b. 5-26. Centrality of Holes. An important appli- cation of the dependent locational tolerance sym- bol to indicate a centrality requirement isinthe dimensioning of holes drilled radially in a cy- lindrical piece. Figure 5-27 shows an implied requirement that the small hole be located ra- dially, that is, the centerline of the small hole will intersect the centerline of the longitudinal hole. However, the requirement is only implied by the drawing, it is not stated. The only stated requirements are that the small hole be 0.200 ±0.002 in diameter, and located longitudinally 0.600 ±0.001 from the seating face. The im- plied requirement is that the small hole be drilled as shown at position 1. However, if the hole is drilled at position 2, it will still meet the stated requirements as to location and di- ameter. When requirements on a drawing are implied instead of stated, the gage engineer must examine the drawings, determine the functions of the various components, and decide the real requirements that will be incorporated in the Figure 5-25. Cocking caused by threads binding before full seating 5-22
gage. It is the desigp engineer, knowing the functions at each component, who should de- cide upon the real requirements, and who should state them on the drawing in the form of sym- bols or notes. If all the requirements are stated and none are implied, the gage engineer can de- sign suitable gages to check the requirements without having any knowledge of the functioning of the components. Figure 5-28 shows the same part, in which the real requirements for the location and size of the small hole are stated. The basic dimension states the theoretically exact longitudinal loca- tion of the center of the hole. The actual center may vary within a tolerance circle, 0.002 in di- ameter around the theoretically exact center, when the hole is produced to the minimum size. 5-27. Basic Angle Dimensioning. This system of dimensioning is explained in Drawing 30-1-7, and no further explanation is given in this sec- tion. However, the engineer should be aware of certain pitfalls in connection with the use of this system. In figure 5, of Drawing 30-1-7, it is important to note that the toleranced dimension 1.500 - 0.005 is specified to sharp corners. This is not a requirement for sharp corners on the part produced. Since MIL-G-2550 specifies WRONG XA.XXX | Figure 5-26. Application of dependent symbol Figure 5-27. Implied requirement for centrality oj small Hole 5-23
Figure 5-28. Correct statement of requirements for small hole that sharp corners shall be broken, the inspector cannot check this dimension directly. The gage engineer will convert the requirements to the da- tum method of dimensioning tapers when desig- nating a gage for such a part. The design en- gineer should also use the datum method wher- ever possible. И sharp corners are required, the note "sharp corners required" should be added to the drawing. 5-28. Effect <Л Gage Tolerance on Component Tolerance. When specifying tolerances on di- mensions, the design engineer must be aware of the effect of gage tolerances on the component tolerances. Although gages are produced by methods that permit closer tolerances than are possible with mass-produced components, there is still atol- e rance which must be allowed to the gage maker. And since the usual type of dimensional gaging requires two gages, GO and NOT GO, the gage tolerances are compounded. The GO gage checks the dimension in the direction of max- imum metal conditions, while the NOT GO gage checks the dimension in the direction of min- imum metal conditions. All gage tolerances are within the component tolerance, and in effect the total amount al the gage tolerances repre- sents a proportion of the component tolerance which is denied the contractor. Tn addition, since parts must enter or be entered by the GO gage to be acceptable, their constant use pro- duces wear on the gaging surfaces. To reduce the expense of frequent replacement of gages, and to ensure that parts will not exceed the maximum permissible metal conditions, a wear allowance, also within the component tolerance, is applied to the GO gage prior to the applica- tion of the gage tolerance. Figure 5-29a shows a simple cylindrical part dimensioned diametrically, the shaded areas Indicating schematically the tolerance zone per- mitted by the drawing. Figure 5-29b shows an enlargement of figure 5-29a, in which there are indicated schematically the effects of Ordnance acceptance gage tolerance and work gsge tol- erances on the tolerance zone permitted by the drawing. It will be noted that there is no wear 5-34
Ul as 8s 1/2 EFFECTIVE TOL. 1/2 1/2 1/2 ORD. GO GAGE WEAR ALLOWANCE ORD.GO GAGE TOLERANCE WORK GO GAGE TOLERANCE TOTAL GO GAGE TOLERANCE 1/2 1/2 EFFECTIVE TOL. i 1/2 TOTAL NOT *-1/2 WORK NOT GO 1/2 ORD. NOT GO GAGE OPERATOR ШДТ FRODVCE WITH* THIS ZONE TO GNSS GAGES PARTS Ш1ГГОЕ ACCEPTED * WITHIN THIS ZWE GY MEASUREMENT GO GAGE TOL.f GAGE TOL. TOL. Figure 5-29. Effect of gage tolerances on component tolerance allowance on the Ordnance NOT GO gage, since parts should not enter or be entered. In addi- tion, any wear will take the gage further Into the tolerance zone permitted by the drawing, and thus give more assurance that nonconforming parts will not be accepted by the gage. It will also be noted that in addition to the encroach- ment on the permitted tolerance sone by Ord- nance gages, work gages normally are dimen- sioned and toleranced within the limits of the Ordnance gages, thus further reducing the tol- erance aone available for the actual producer of the parts. However, no wear allowance, as such, is applied on the work GO gages.
ENGINEERING DESIGN HANDBOOK SERIES Listed below ere the Handbook! which have bean published or submitted for publication. Handbooks with publica- tion dates prior to 1 August 1962 wars published as 20-seriee Ordnance Corps pamphlets. AMC Circular 310-38, 19 July 1963, redesignated those publications as 706-series AMC pamphlets (i.s., ORDP 20-138 was redesignated AMCP 706-138). All now, reprinted, or revised Handbooks are being published as 706-series AMC pamphlets. General and Miscellaneous Subjects No. Tide 106 Elements of Armament Engineering. Part One, Sources of Energy 107 Elements of Armament Engineering, Part Two, Ballistics 100 Elements of Armament Engineering, Part Three, Weapon Systems and Componsnts 110 Experimental Statistics, Section 1, Basic Con- cepts and Analysis of Measurement Data 111 Experimental Statistics, Section 2, Analysis of Enumerative and Claseificatory Data 112 Experimental Sts tie ties, Section 3, Planning and Analysis of Comparative Experiments 113 Experimental Statistics, Section 4, Special Topics 114 Experimental Statistics, Section 5, Tables 121 Packaging and Pack Engineering 134 Maintenance Engineering Guido for Ordnance Design 135 Inventions, Patents, and Related Matters (Revised) 136 Servomechanisms, Section 1, Theory 13? Servomechanisms, Section 2. Msasur ament and Signal Converters 13S Servomechanisms, Section 3, Amplification 139 Servomechanisms, Section 4, Power Elements and System Desipi 170(G) Armor and Its Application to Vehicles (U) 250 Guns --General (Guns Series) 252 Gun Tubss (Guns Series) 270 Propellant Actuated Devices 290(C) Warheada--General (U) 331 Compensating Elements (Fire Control Series) 355 The Automotive Assembly (Automotive Seriss) (Revised) Ammunitisn and Explosive. Series 175 Solid Propellants, PartOns 176(C) Solid Propellants, Part Two (U) 177 Propertiss of Explosives of Military Interest, Section 1 178(C) Properties of Explosives of Military Interest, Section 2 (U) 179 Explcsivs Trains 210 Fuses, General and Mechanical 211(C) Fuses, Proximity, Electrical, Part One (U) 212(3) Fusee, Proximity, Electrical. Part Two (U) 213(8) Fuses, Proximity, Electrical, Part Three (U) 214(S) Fuses, Proximity. Electrical Part Four (U) 215(C) Fuses, Proximity, Electrical. Part Five (U) 244 Section I. Artillery Ammunition--General, with Table of Contente, Glossary and Index for Series 245(C) Section 2, Design for Terminal Effects (U) 246 Section 3, Design for Control of Flight Characteristics 247 Section 4, Design for Projection 248 Section 5, Inspection Aspscts of Artillery Ammunition Design 249 Section 6, Manufacture of Metallic Components ' of Artillery Ammunition Ballistic Missile Series No. Title 281(B-RD) Weapon System Effectiveness (U) 282 Propulsion and Propellants 283 Aerodynamics 284(C) Trajectories (U) 286 Structures Ballistics Series 140 Trajectories, Differential Effects, and Data for Projectiles 150 Interior Ballistics of Gone 160(S) Elements of Terminal Ballistics, Part One, Introduction, Kill Mechanisms, and Vulnerability (U) 161(3) Elements of Terminal B^allistics, Part Two, Collection and Analysis of Data Concerning Targets (U) 162(8-RD) Elements of Terminal Ballistics, Part Three, Application to Missile and Space Targets (U) Carriage* i and Mounts Series 340 Carriages and Mounts—General 341 Cradles 342 Recoil Systems 343 Top Carriages 344 Bottom Carriages 345 Equilibrators 346 Elevating Mechanisms 347 Traversing Mechanisms Military Pyrotechnics Series 1»6 Part Two, Safety, Procedures and Glossary 187 Part Three. Properties of Materials Used in Pyrotechnic Compositions Surfac»to-Air Missile Series 291 Part One, System Integration 292 Part Two, Weapon Control 293 Part Three, Computers 294(3) Part Four, Missile Armament (U) 295(d) Part Five, Countermeasures (U) 296 Part Six, Structures and Power Sources 297(8) Part Seven, Sample Problem (U) Materials Series» 149 Rubber and Rubber-Like Materials 212 Gasket Materials (Nonmetallic) 691 Adhesives 692 Guide to Selection of Rubber O-Rings 693 Magnesium and Magnesium Alloys 694 Aluminum and Aluminum Alloys 697 Titanium and Titanium Allots 698 Copper and Copper Aiioye 699 Guide to Specifications for Flexible Rubbe Products 700 Plastics 721 Corrosion and Corrosion Protection of Metals 7 22 Glass •The Materials Series is being published as Military Handbooks (MIL-HDBK-) which are available to Department of Defense Agencies from the Naval Supply Depot, 5801 Tabor Avenue. Philadelphia. Pennsylvania 19120.